AI-Driven Instruction Interfaced with Task-Based Learning in Grade 8 Mathematics

Abstract

Modern secondary mathematics education faces a procedural gap between basic calculations and higher-order statistical literacy. This study evaluated the integration of artificial intelligence tools (ChatGPT and Dola AI) with task-based learning (TBL) to enhance student performance. Adopting a mixed-methods design with a one-group pretest-posttest experimental approach, the study analyzed data from 34 Grade 8 students in Abra, Philippines. Data were gathered using DepEd Project SMART-adapted tests validated by experts with a 4.97 content validity index alongside engagement surveys and interviews. Quantitative data were analyzed via weighted means and paired t-tests, while interviews captured qualitative challenges. Baseline results showed a "Fairly Satisfactory" performance (Overall M = 76.89) and a procedural ceiling in standard deviation calculation (M = 68.71, "Did Not Meet Expectation"). Post-intervention, Project-Based Learning (PBL) achieved the highest performance (Overall M = 85.18, "Very Satisfactory"), followed by Think-Pair-Share (M = 83.18) and Jigsaw (M = 80.04). Paired t-tests confirmed significant gains across all frameworks (p < 0.01). PBL yielded the largest overall improvement (Gain = 8.29, t = 13.86) and standard deviation breakthrough (Gain = 5.50, t = 9.09), while Think-Pair-Share maximized range growth (Gain = 10.21, t = 7.45). The intervention earned a "Very High Level of Perceived Effectiveness" (M = 3.43), driven by task practicality (M = 3.62). In conclusion, combining AI with TBL shifts classrooms into active environments, framing math as a practical tool. However, it does not automatically guarantee full technical proficiency. While PBL is the most powerful strategy for complex problem-solving, operational benefits are heavily constrained by weak school Wi-Fi, algorithmic inaccuracies, and cognitive overload. To achieve sustainability, institutions must pair these pedagogical models with reliable digital infrastructure, technical support, and clear instructional guidelines that foster responsible AI literacy.